I've had an EQ account for probably 5 years. I have hardly played it in the last year... something about being in Australia and never being able to find a group mature enough to really achieve something with at the times I could be on, never mind that prime time down-under seems to regularly coincide with patch time. I've been playing it on a 3GHz P4 with 1G RAM, and an nVidia FX5600 card, using a 512/128k ADSL connection, and when there are over a hundred PCs within the field of vision, the flame rate gets very ordinary. I believe my current installation of EQ is chewing a few Gb of hard drive space too, but I allocated 5G to that, so it's ok for now...
The opportunity came up for me to play on the last beta weekend, reported a few bugs and annoyances, and although there are a few things I don't agree with, I was most impressed.
Things I like about it:
Camping disputes are eliminated. So too is training on other parties/individuals. When you go on a mission, the group of people there is your choice. Your instance of the zone is essentially yours and yours alone!
If you can't get a group, you can "recruit" a bunch of NPCs, which do the job of their class pretty well. If you die, your NPC monk will rez you once things settle down, and your pet (a permanently charmed animal) will fight on to the death too, instead of clearing off without its master! This is handy when you will be rezzed later.
Death does not impose an unconditional many-hours-worth-of-experience penalty on you. It doesn't really matter if you die during a mission, but you effectively suffer a morale type skills penalty until you kill enough "big enough" things to get it back. And every key figure killed during the mission can net you a "morale bonus" of sorts. This is reset when you leave the mission/area - so you can abort the mission and start again to recover from a penalty, or complete the final task using skill, strength and just a bit of luck.
The graphics engine is an extremely cool, efficient and tight piece of work. It renders everything beautifully. I love the way that the capes flutter in such a realistic way with character movement, even when your ranger is breakdancing.
If you're standing on higher ground, your arrows fly further and hence you shoot targets further away. Also when you click to attack, it will try to take you to the outside range before letting your first shot off, instead of simply telling you you're out of range and letting you do the guesswork.
Pathing has had a lot of work done on it. If you can see a spot on the other side of the valley, but the way there isn't at all obvious, then you have to find the way yourself. But clicking on a point on the ground beneath the current ledge takes you there in a sensibly efficient manner, taking you on a path around obstacles instead of aiming through them. This also means your pet and/or recruited NPCs will not lose sight of you. If they die, well that's another matter.
No more loot disputes. By yourself, if you kill it, the loot's yours. In a group, the system spreads the loot around fairly and highlights stuff assigned to you for pickup.
There is a really nice quest log that stores where you're up to for all outstanding quests. Brilliant! There's also usually a helpful arrow on the local terrain map pointing toward your destination. The struggle is getting through/past the obstacles, not getting lost.
The local terrain overhead map is clickable. You can indicate where things are by clicking on points, or drawing lines (which fade) to show which direction your party should go. Call it a map equivalent of hand-signals.
I got to fire a trebuchet! Repeatedly! No, there was no experience given for the multiple kills achieved, but had our party walked into that we would have been decimated...
Being surrounded by a horde of little guys is inherently dangerous, no matter if you're level 10 and they're all level 3. However, killing them will still net
Ummm, I understand that breeding may be important, but what part do females play in that??? How does it work??? Are they important? Surely not as important as computers!
I was convinced on Linux stability during the 1.2 kernel series, when I went on a holiday, leaving my box up and running, the X screensaver ticking over...
2 weeks later, I got back and the box was slow. Extremely slow. But I (eventually) got in, and managed to have a look around. The screensaver had a memory leak, and was using 40Mb of memory - virtually all 16Mb of RAM, plus 24 of my 32Mb swap partition, thrashing constantly. I killed the screensaver and within seconds everything was returning to normal. No reboot required, no nothing.
In my work, I have supported from 120 - 180 boxen in the field and running the 2.4 series kernel - they get used most days of the year, and they only restart due to one of the following.
Some bozo accidentally hitting the power button.
Some bozo accidentally pulling the power cable.
Some bozo accidentally turning off the UPS.
Some bozo not plugging the box into a UPS (and having a short power outage kill the box).
Some bozo causing an extended power outage.
Hardware failure - typically when we need to replace a failing hard drive (in a software mirrored pair) - and data loss has not been experienced on any of them in the 3 years since I started working with them.
Recently I noted one of our boxes had an uptime of just over 500 days, and it had copped something like 15 software/config changes in that time. 200 days uptime is nothing special except to note that nobody did anything stupid within the vicinity of the machine in that time!
So from my experience, if you run a stable kernel then your biggest threats are physical, whether they be carbon or silicon based.
Oookay, now I'm going to go off my rocker a bit. No proof, no sums, no drugs, but...
What if a "black hole" was simply a gravity well powered particle accelerator of the atom smashing variety? Possible? I dunno. Is there a possibility that there's just enough matter in your black hole to accelerate matter to 0.999c and collide with the surface of what was once a neutron star (in the presence of trapped photons), creating nothing but subatomic particles (dark matter???) and energy?
To me, throwing off lots of energy might help explain this universal expansion thing. Then again, I could be barking up completely the wrong eucalyptus. An interesting thought experiment nontheless. Any theories/reasons against it, or should I just sign my self into the institution now?
No no no! You never changed your mind! The product delivered simply didn't meet your expectations, therefore they failed to properly analyse the requirements and determine that "coffee" really means "tea" and you were absolutely certain that the "bagel" feature is a well established standard that should have been on by default!!!
Never mind that immediately after ordering the product you relocated yourself to another part of the building, fully expecting the delivery address to be updated on-the-fly.
Enforcing referential integrity constraints can certainly slow things down, but I've always found it far preferable to data inconsistency. When I load a log file into a database I want to know that all or nothing arrived, then there's no screwing around trying to feed half the file back in later. This is never a problem with a full transactional RDBMS.
But you can also get around speed issues when you are sure of the data. Years ago I was an acting Postgres DBA - for a couple of years or so - and quickly discovered that to speed up data dump/load issues it was a relatively simple task to disable referential integrity enforcement before the load, and reenable afterward. This was especially useful when my manager accidentally messed up a single client data table with a single update and I was able to disable the table's triggers, drop it, load the morning's backup, and reenable the integrity enforcement.
Indeed, a database restore is typically of the order: Create the tables, load the data, turn on integrity. ~10G in 15 minutes (backup and restore) on now-5-year-old-hardware, not hard to do.
Oracle offers similar loading functionality. I've heard a very highly regarded DBA referring to it as ripping a hole in the side of the table, but basically it's a referential integrity bypass. Great stuff if you know what you're doing, and it can turn hours of loading into something better measured in minutes.
But only if they know at least some DB basics! Sometimes it is better that they at least run the query by someone familiar with how to write a proper query. Then, when they come to you screaming that the database is slow, and you see:
SELECT * from foo,foo,foo,foo,foo,foo,foo...
Then you're perfectly within your rights to "retrain" them. With pliers.
Australian research has turned up something "promising" using the drug Clioquinol - not without unpleasant side-effects, but chosen because it has already been used for other things (maybe scientists can come up with something friendlier). More about its initial promise on this story from Catalyst - the science show on Australia's non-commercial channel, the ABC. (Includes links to watch the story with Real or Windows Media players.)
As far as WTF australia is thinking, they have consistantly not gotten it when it comes to the internet...
First consider that (maybe I am being a bit simplistic here) federal legislation starts with the government. Then look at this picture. Then consider that the federal leader is the short guy left of the middle up the front. Yes, the funny-looking one who doesn't quite qualify to enter a serious baldness contest. Yes, he's the captain of the team pictured.
But if I live on the dole, or some other government welfare, it would be ok? Actually, considering the costs involved in dealing with these situations, it might still take a while for the government to break even.
So all you no-hopers considering careers of living off the tax payer, please do us a favour - if you ever decide to commit suicide, please do it early on as opposed to draining society for 20 years first. I'll not tell you how to do it though, because that will surely get me in trouble and I don't have that kinda money on me.
If everyone followed Larson then I'd be be a lot more paranoid than just the tin-foil hat! Beware the poker-playing critters.
Anyway, smite, smote, smitten... yes indeed 'tis a stoopid language. It's soooo much fun explaining that there are so many exceptions to the spelling rules, even to my rather bright 7-year-old.
But back to the only slightly OT. Indeed, any non-denominational-specific god could have certainly found a way to smu^Ho^Hite Redmond if they really wanted to, unless it's protected as some sort of nexus for the (closed-)source of all evil in the world?:-P
Hmmm... firstly, although mathematics is indeed a very precise language, it still fails to define the number 1. That kinda requires a reference point which is not circularly linked with itself.
Secondly - and I'm being very hypothetical here - even though dimensions are implied to be static, surely a reference point within one dimension can move independently of other dimensions? And aren't our observations based on drift relative to the reference point being used?
If I accelerate (and thus alter movement) upwards, then that is relative to where I was - but if I'm bouncing up and down in a moving bus, or indeed, on a planet hurtling through space, I am surely still accelerating relative along the Z axis - with reference to the bus, and only really accelerating along one axis relative to the bus, the road, the sign post, whether or not I was already moving in relation to it or not. And if the bus accelerates while I'm in the air, then it moves forward relative to me or I move backwards relative to it - taking the end with the head lights as the bus' front.
To me, "rotation in 1 dimension" is possible, with a very limited definition of rotation - freedom to change "forward" from a given direction to its opposite. If anyone could define the "1 dimension" part more precisely (of which I believe the hard part is defining "1"), I'd be very interested in the result. But let's assume that rotation means a change in vector, which requires at least 2 dimensions of freedom, and at least one more dimension, so the first two have something to change in.
The fun part is being able to accelerate through time - using relativity theory, or whatever, while moving along a given physical axis. If I can alter a the vector of an object which has freedom in 2 dimensions, surely that's rotation. So if I can alter the vector of an electron with the dimensions being one physical axis, and time, isn't that rotation in the traditional 2 dimensional sense too? Us three-dimensional beings have the luxury of freedom of accelerating in a whole three dimensions - relative to the reference dimension of time. The next factor is that in order to be able to move through the 4th dimension, then there must be a 5th, whose outcome we cannot both determine and measure at the same time. I thought that dimension was probability (or has it been redefined since I last read about it - it's been a while), and this whole probability thing is what quantum mechanics gets all funky with.
It seems to me we're talking about movement through time and space relative to probability. Does this make sense?
But seriously (just a little OT), the response to a knock can be tuned easily enough:
Firewall. Your bouncer only lets in whoever he's been taught to trust. Or you can give it a guest list. Many broadband interfaces can also present a "false" front door thanks to IP Masquerading. Neither is 100% foolproof, but they do make life harder, especially for bulk tools used by script kiddies.
Silently DROP incoming SYN packets on unused ports. Like having a trapdoor under the doormat - what knock?
Something I liken to Neighbourhood Watch - at the first sign of a port scanner, broadcast to your friends and concerned neighbours of the attempt so they'll be wary of the stranger.
Use your own bot army to DoS the attempted intruder. Something like a Claymore on the doorstep?
Then there's antivirus, groupware... the difference as I see it is the tools to do these are freely available with basically anything *n[iu]x*, while you tend to have to pay for a decent solution that runs on your favourite monopolistic vendor's OS. Not always, mind, but typically. Since I payed for XP (keeping it up to date), no software but games have cost me anything - AVG/OpenOffice/Mozilla + extensions/software that comes with purchased hardware... etc etc... it's pretty easy to meet license terms when you're not putting things to commercial use. This also means I'm not running any networked services publicly, so this box never accepts an incoming connection from the cloud.
As for the stuff that does matter - web, database etc services... I leave that to my Linux box, running just what it needs to, and I take a little time semi-regularly to ensure it stays close enough to up-to-date. It hasn't let me down as yet (neither did FreeBSD while I was running that too), and this is year 13...
Disclaimer: I don't know everything, but I know what ideas I like. And just because I like the idea, doesn't necessarily mean I implement it.
I remember, about 10 years ago (give or take - 12 maybe?), evaluating FOSS proxy caches for a company as part of a short-term contract. One was Squid, another - bit of an upstart - was Apache in proxy mode. I was not very impressed with Apache at the time. I thought its storage methods were silly, and it had a lot of optimisation to do before it could even think of going anywhere.
I guess we're here now, and we probably have been for some time - but that appears to have quietly slipped in while I wasn't looking.
Then I moved into a position with a company selling a solution with PAID FOR LICENSES of Netscape included. We were happy to pay the fee though, because it did things for us that simply didn't work otherwise on Windows 3.1(/1) - the choice of clients of my old employers...
Now, although I thought those large warrior women were around a bit longer ago than 10 years, at least I know what they are... but what's an "ebay"??
(Disclaimer: This is my personal opinion.) If you're building your own box for your own purposes, and expect to be able to fix things yourself based on google/forums/friends, then don't go RedHat. It's too limited in scope for that. FC gets good support from their development framework, but again you don't need it. From what I see on a daily basis, RedHat's big plus is it's heavily certified with Oracle (and I'm not sure what else, because that really doesn't concern me in my work). There's only a few distros that have this support advantage, and RedHat was one of the first there. I'm fond of SuSE myself, but we can't justify going that way with the local support we can get if we have to.
As for making the jump from EL3 to EL4, well the main reasons IMHO are to dump all the backported patches made since EL3's inception first, going with packages a little less off-the-beaten-track, and then a few updates of things that help the job for frustrated admins. Little things like installing on logical volumes at the outset (long overdue!) and the nature of LVM 2, which allows taking multiple read-writable snapshots of any logical volume, and if lvcreate's usage is to believed, at some point we will be able to take snapshots of snapshots.
By far RedHat's biggest failing IMO is the lack of support for ReiserFS - JFS and XFS would be nice for others, but the former is all I really care for. I like having a filesystem that genuinely allows for atomic disk transactions without any noticeable performance hit. But as has already been stated, RedHat aren't interested in supporting it. It's a real shame, but something we have to live with for now.
Indeed, true AI is in a program that "instinctively" knows how to learn. When we can give it an "environment", and its "skills" can be based on trial and error, learning by "observation" and "experience", then we have something truly amazing. I guess the key is figuring out how to evolve learning abilities in the organisms, and giving them some sort of interface which they may "discover", so they may impose modifications on their environment, tempered by some basic "laws of nature".
This includes perhaps migrational issues (making it hard to get from one processor of the cluster to another, as if there was a mountain or an ocean between them), resource limitation and even fluctuation (as in introducing a resource drought), so survival depends on versatility, ruthlessness or being solely dependent on the nature of environmental variables. Domination by figuring out and culling competitors for resources, or even feeding on them somehow would be fun to see, but it would have to be allowed/encouraged to evolve.
Also "natural disasters" can be introduced, which might make some species extinct, or if you're lucky, force an evolutionary leap. This could be achieved by throwing some randomness into the works - maybe semi-frequent but minor or localised events, occasionally something extreme. Applying bitwise mask - either one - to a section of their memory with a bit sequence containing some density of bits depending on the severity of the event perhaps, i.e. everything from minor earthquakes applying an XOR mask mostly 0's to a chunk of a node's memory (which won't change much, but might do something significant) to an organism driven nuclear catastrophe type event (which ANDs most of the environment with mostly 0's)?
It would be really cool if these "organisms" could peek at the nature of their environment, and develop survival instincts based on event prediction. Like "last time we saw those three operations in a row, we should have left town!"
I imagine this would become fairly complex fairly quickly, but would be great fun to play with. Imagine having a different type of environmental setup on each cluster node. Bruahahaha, let's change the nature of all of the cluster's nodes suddenly - switching many, losing a few and gaining a few never-before-seen.
Then at what point do we go "awww those little operations are just doing the darndest things that I can't possibly turn off the PC now..."? Or better still they find a fragment of memory containing a dictionary "artifact" and one day you see they've "broadcast" a message to the universe on your console:
Looking here, I see lots of what I would normally consider insightful input into what could be eventual definitions of stars, supergiants, planets, asteroids and pebbles. However, as was mentioned the definitions are simply going to have to evolve with our understanding.
Would The Earth cease to be a planet just because something threw it forever out of our solar system? (Well actually, for now almost certainly yes, 'cos then there'd be no humans to define "planet";). At what point does an asteroid have to collect enough dust and become spherical enough to become a planet? Not all planets are spherical - Mercury is more elliptical from memory, thanks to effects of being in close proximity to a star... errr, the Sun. They wouldn't even have to necessarily spin - though that would help with roundness.
Also from recollection of earlier dictionaries, our moon would become a planet (or planetoid?) if some catastrophe yanked it away from the earth, to forever go around the sun - because it wouldn't then be a body orbiting a planet - a simple, but rock solid definition IMHO. Oh but hang on, what about all those little rocks orbiting Earth???
In that respect remember that some definitions are probably inherently transitional, depending on what they are doing. If it's a rock orbiting a sun, it's an asteroid, around a planet then it's a moon, if it's become round (has enough gravitational pull to hold itself together?) then it's a planet if it's going around a Sun - or is it, because what if the planet escapes?
I believe the dictionary definition of "moon" is pretty good already, but as for the rest... I hope you can see what I mean because it gives me a headache! If we set a strict definition of a type of celestial body, and then suddenly we discover that there are so many more bodies that just don't quite fit the category, then what? I don't fancy taking liquid paper to my dictionary. So I will leave splitting those hairs to the experts.
I do mean almost. The www wasn't exactly a phenomenon then, but using integers to represent coordinates has always been a necessity if you are wont for floating point numbers in your system. Especially if you're working with a limited amount of processing power, or the cost of simulating operations gets relatively high.
Summary: I was working on an LPMud which was darn well going to have a coordinate system for all "rooms" such that it would easily calculate who was within a certain radius of any action. I grabbed my old maths book, did the trigonometry, came up with the relevant equations, boiled them down, and then proceeded to implement them in the day's game driver. The game driver had NO support for floating point operations, so I used a few tricks:
Refer to coordinates in degrees, but increased a couple of orders of magnitude to make fractions of coordinates representable (store a precise value as an acceptably less precise integer). I even scaled it - 800 "degrees" instead of 360 seemed to work best for my needs.
Implement simple series for some trigonometric calculations - I got sufficiently accurate values with IIRC the first 5 steps in the series
Where speed really was an issue, I used incomplete trigonometric lookup tables, with some allowance in the functions to approximate results as "x/yths between a[z] and a[z+1]."
Bingo, one working coordinates scheme to accomodate the physics I wanted to allow for, with absolutely no real numbers in play.
I did not even think of patenting it. Why? Because it only took an afternoon's worth of thought to do all the necessary derivations and calculations, and it was all common sense! Besides, I might have had to have credited my high school math text!;) The only real difference is I was representing the coordinates in 2 separate variables, instead of concatenating them with a comma. But then I wasn't trying to store them in a URL.
The implementation of all the basics in LPC took me probably the better part of a week, and was tested, and worked to an acceptable degree, but that was just coding. It's just a shame I went on to start University before I could release it and give it momentum to grow.
[rant]Grow up Microsoft, and patent something that a high school student can't do.[/rant]
Users predictable if you let them get away with it
on
Password Security Panned
·
· Score: 2, Interesting
Probably 10 years ago I went to a security talk, which mentioned a few passwords that users seem to like using, and always seems to get picked up by apps like crack... at the time, on the ISP I was using there were no rules for the customers' passwords - why make life hard for paying customers anyway? About a week later, I was logged onto the SCO Xenix box at my ISP, which got someone else's (UUCP feed) dialup TTY mixed up with mine, and dumped a copy of all their traffic to my session. I saw their login and password, and a copy of their data stream; The password was one of the top few mentioned at the talk. No surprises: mypass.
Unfortunately, if we don't have complex "Don't start with a number, the new one must not be similar to the last, do this, don't do that" rules, users will tend to take the easy way out and use "password" if given the option. It seems today that the only way to ensure something random is to reduce the number of allowable permutations. Dictionary cracks become meaningless when the user has no statistical preference for leaning on dictionary words. Given the choice, I would just as likely use "A2jj*Z,L" as "dictionary" for a password, but Joe Average goes and spoils it...
Slashdot Mirror
The opportunity came up for me to play on the last beta weekend, reported a few bugs and annoyances, and although there are a few things I don't agree with, I was most impressed.
Things I like about it:
Ummm, I understand that breeding may be important, but what part do females play in that??? How does it work??? Are they important? Surely not as important as computers!
Yours confusedly,
Errr...
2 weeks later, I got back and the box was slow. Extremely slow. But I (eventually) got in, and managed to have a look around. The screensaver had a memory leak, and was using 40Mb of memory - virtually all 16Mb of RAM, plus 24 of my 32Mb swap partition, thrashing constantly. I killed the screensaver and within seconds everything was returning to normal. No reboot required, no nothing.
In my work, I have supported from 120 - 180 boxen in the field and running the 2.4 series kernel - they get used most days of the year, and they only restart due to one of the following.
- Some bozo accidentally hitting the power button.
- Some bozo accidentally pulling the power cable.
- Some bozo accidentally turning off the UPS.
- Some bozo not plugging the box into a UPS (and having a short power outage kill the box).
- Some bozo causing an extended power outage.
- Hardware failure - typically when we need to replace a failing hard drive (in a software mirrored pair) - and data loss has not been experienced on any of them in the 3 years since I started working with them.
Recently I noted one of our boxes had an uptime of just over 500 days, and it had copped something like 15 software/config changes in that time. 200 days uptime is nothing special except to note that nobody did anything stupid within the vicinity of the machine in that time!So from my experience, if you run a stable kernel then your biggest threats are physical, whether they be carbon or silicon based.
What if a "black hole" was simply a gravity well powered particle accelerator of the atom smashing variety? Possible? I dunno. Is there a possibility that there's just enough matter in your black hole to accelerate matter to 0.999c and collide with the surface of what was once a neutron star (in the presence of trapped photons), creating nothing but subatomic particles (dark matter???) and energy?
To me, throwing off lots of energy might help explain this universal expansion thing. Then again, I could be barking up completely the wrong eucalyptus. An interesting thought experiment nontheless. Any theories/reasons against it, or should I just sign my self into the institution now?
Never mind that immediately after ordering the product you relocated yourself to another part of the building, fully expecting the delivery address to be updated on-the-fly.
Hmmm... have they fixed these yet though?
Enforcing referential integrity constraints can certainly slow things down, but I've always found it far preferable to data inconsistency. When I load a log file into a database I want to know that all or nothing arrived, then there's no screwing around trying to feed half the file back in later. This is never a problem with a full transactional RDBMS.
But you can also get around speed issues when you are sure of the data. Years ago I was an acting Postgres DBA - for a couple of years or so - and quickly discovered that to speed up data dump/load issues it was a relatively simple task to disable referential integrity enforcement before the load, and reenable afterward. This was especially useful when my manager accidentally messed up a single client data table with a single update and I was able to disable the table's triggers, drop it, load the morning's backup, and reenable the integrity enforcement.
Indeed, a database restore is typically of the order: Create the tables, load the data, turn on integrity. ~10G in 15 minutes (backup and restore) on now-5-year-old-hardware, not hard to do.
Oracle offers similar loading functionality. I've heard a very highly regarded DBA referring to it as ripping a hole in the side of the table, but basically it's a referential integrity bypass. Great stuff if you know what you're doing, and it can turn hours of loading into something better measured in minutes.
SELECT * from foo,foo,foo,foo,foo,foo,foo ...
Then you're perfectly within your rights to "retrain" them. With pliers.
Maybe something will come of it.
First consider that (maybe I am being a bit simplistic here) federal legislation starts with the government. Then look at this picture. Then consider that the federal leader is the short guy left of the middle up the front. Yes, the funny-looking one who doesn't quite qualify to enter a serious baldness contest. Yes, he's the captain of the team pictured.
Would you expect anything different?
So all you no-hopers considering careers of living off the tax payer, please do us a favour - if you ever decide to commit suicide, please do it early on as opposed to draining society for 20 years first. I'll not tell you how to do it though, because that will surely get me in trouble and I don't have that kinda money on me.
If everyone followed Larson then I'd be be a lot more paranoid than just the tin-foil hat! Beware the poker-playing critters.
Anyway, smite, smote, smitten... yes indeed 'tis a stoopid language. It's soooo much fun explaining that there are so many exceptions to the spelling rules, even to my rather bright 7-year-old.
But back to the only slightly OT. Indeed, any non-denominational-specific god could have certainly found a way to smu^Ho^Hite Redmond if they really wanted to, unless it's protected as some sort of nexus for the (closed-)source of all evil in the world? :-P
I believe you're looking for smote. Stoopid language...
A lesson for the folks at home - don't play Russian Roulette with anything other than a revolver.
But then you must define set theory...
Secondly - and I'm being very hypothetical here - even though dimensions are implied to be static, surely a reference point within one dimension can move independently of other dimensions? And aren't our observations based on drift relative to the reference point being used?
If I accelerate (and thus alter movement) upwards, then that is relative to where I was - but if I'm bouncing up and down in a moving bus, or indeed, on a planet hurtling through space, I am surely still accelerating relative along the Z axis - with reference to the bus, and only really accelerating along one axis relative to the bus, the road, the sign post, whether or not I was already moving in relation to it or not. And if the bus accelerates while I'm in the air, then it moves forward relative to me or I move backwards relative to it - taking the end with the head lights as the bus' front.
To me, "rotation in 1 dimension" is possible, with a very limited definition of rotation - freedom to change "forward" from a given direction to its opposite. If anyone could define the "1 dimension" part more precisely (of which I believe the hard part is defining "1"), I'd be very interested in the result. But let's assume that rotation means a change in vector, which requires at least 2 dimensions of freedom, and at least one more dimension, so the first two have something to change in.
The fun part is being able to accelerate through time - using relativity theory, or whatever, while moving along a given physical axis. If I can alter a the vector of an object which has freedom in 2 dimensions, surely that's rotation. So if I can alter the vector of an electron with the dimensions being one physical axis, and time, isn't that rotation in the traditional 2 dimensional sense too? Us three-dimensional beings have the luxury of freedom of accelerating in a whole three dimensions - relative to the reference dimension of time. The next factor is that in order to be able to move through the 4th dimension, then there must be a 5th, whose outcome we cannot both determine and measure at the same time. I thought that dimension was probability (or has it been redefined since I last read about it - it's been a while), and this whole probability thing is what quantum mechanics gets all funky with.
It seems to me we're talking about movement through time and space relative to probability. Does this make sense?
Mmmm... sentry guns.
But seriously (just a little OT), the response to a knock can be tuned easily enough:
- Firewall. Your bouncer only lets in whoever he's been taught to trust. Or you can give it a guest list. Many broadband interfaces can also present a "false" front door thanks to IP Masquerading. Neither is 100% foolproof, but they do make life harder, especially for bulk tools used by script kiddies.
- Silently DROP incoming SYN packets on unused ports. Like having a trapdoor under the doormat - what knock?
- Something I liken to Neighbourhood Watch - at the first sign of a port scanner, broadcast to your friends and concerned neighbours of the attempt so they'll be wary of the stranger.
- Use your own bot army to DoS the attempted intruder. Something like a Claymore on the doorstep?
Then there's antivirus, groupware... the difference as I see it is the tools to do these are freely available with basically anything *n[iu]x*, while you tend to have to pay for a decent solution that runs on your favourite monopolistic vendor's OS. Not always, mind, but typically. Since I payed for XP (keeping it up to date), no software but games have cost me anything - AVG/OpenOffice/Mozilla + extensions/software that comes with purchased hardware... etc etc... it's pretty easy to meet license terms when you're not putting things to commercial use. This also means I'm not running any networked services publicly, so this box never accepts an incoming connection from the cloud.As for the stuff that does matter - web, database etc services... I leave that to my Linux box, running just what it needs to, and I take a little time semi-regularly to ensure it stays close enough to up-to-date. It hasn't let me down as yet (neither did FreeBSD while I was running that too), and this is year 13...
Disclaimer: I don't know everything, but I know what ideas I like. And just because I like the idea, doesn't necessarily mean I implement it.
The library?
I guess we're here now, and we probably have been for some time - but that appears to have quietly slipped in while I wasn't looking.
Then I moved into a position with a company selling a solution with PAID FOR LICENSES of Netscape included. We were happy to pay the fee though, because it did things for us that simply didn't work otherwise on Windows 3.1(/1) - the choice of clients of my old employers...
Now, although I thought those large warrior women were around a bit longer ago than 10 years, at least I know what they are... but what's an "ebay"??
Indeed we don't, not for the database files at least. However, it's hard to go past for other filesystems, including Oracle's apps.
As for making the jump from EL3 to EL4, well the main reasons IMHO are to dump all the backported patches made since EL3's inception first, going with packages a little less off-the-beaten-track, and then a few updates of things that help the job for frustrated admins. Little things like installing on logical volumes at the outset (long overdue!) and the nature of LVM 2, which allows taking multiple read-writable snapshots of any logical volume, and if lvcreate's usage is to believed, at some point we will be able to take snapshots of snapshots.
By far RedHat's biggest failing IMO is the lack of support for ReiserFS - JFS and XFS would be nice for others, but the former is all I really care for. I like having a filesystem that genuinely allows for atomic disk transactions without any noticeable performance hit. But as has already been stated, RedHat aren't interested in supporting it. It's a real shame, but something we have to live with for now.
This includes perhaps migrational issues (making it hard to get from one processor of the cluster to another, as if there was a mountain or an ocean between them), resource limitation and even fluctuation (as in introducing a resource drought), so survival depends on versatility, ruthlessness or being solely dependent on the nature of environmental variables. Domination by figuring out and culling competitors for resources, or even feeding on them somehow would be fun to see, but it would have to be allowed/encouraged to evolve.
Also "natural disasters" can be introduced, which might make some species extinct, or if you're lucky, force an evolutionary leap. This could be achieved by throwing some randomness into the works - maybe semi-frequent but minor or localised events, occasionally something extreme. Applying bitwise mask - either one - to a section of their memory with a bit sequence containing some density of bits depending on the severity of the event perhaps, i.e. everything from minor earthquakes applying an XOR mask mostly 0's to a chunk of a node's memory (which won't change much, but might do something significant) to an organism driven nuclear catastrophe type event (which ANDs most of the environment with mostly 0's)?
It would be really cool if these "organisms" could peek at the nature of their environment, and develop survival instincts based on event prediction. Like "last time we saw those three operations in a row, we should have left town!"
I imagine this would become fairly complex fairly quickly, but would be great fun to play with. Imagine having a different type of environmental setup on each cluster node. Bruahahaha, let's change the nature of all of the cluster's nodes suddenly - switching many, losing a few and gaining a few never-before-seen.
Then at what point do we go "awww those little operations are just doing the darndest things that I can't possibly turn off the PC now..."? Or better still they find a fragment of memory containing a dictionary "artifact" and one day you see they've "broadcast" a message to the universe on your console:
Hello, God is that you?
Is that when you start wishing you bought a UPS?
In Australia, some call it bush bashing. Others call it four wheel driving!
Would The Earth cease to be a planet just because something threw it forever out of our solar system? (Well actually, for now almost certainly yes, 'cos then there'd be no humans to define "planet" ;). At what point does an asteroid have to collect enough dust and become spherical enough to become a planet? Not all planets are spherical - Mercury is more elliptical from memory, thanks to effects of being in close proximity to a star... errr, the Sun. They wouldn't even have to necessarily spin - though that would help with roundness.
Also from recollection of earlier dictionaries, our moon would become a planet (or planetoid?) if some catastrophe yanked it away from the earth, to forever go around the sun - because it wouldn't then be a body orbiting a planet - a simple, but rock solid definition IMHO. Oh but hang on, what about all those little rocks orbiting Earth???
In that respect remember that some definitions are probably inherently transitional, depending on what they are doing. If it's a rock orbiting a sun, it's an asteroid, around a planet then it's a moon, if it's become round (has enough gravitational pull to hold itself together?) then it's a planet if it's going around a Sun - or is it, because what if the planet escapes?
I believe the dictionary definition of "moon" is pretty good already, but as for the rest... I hope you can see what I mean because it gives me a headache! If we set a strict definition of a type of celestial body, and then suddenly we discover that there are so many more bodies that just don't quite fit the category, then what? I don't fancy taking liquid paper to my dictionary. So I will leave splitting those hairs to the experts.
Summary: I was working on an LPMud which was darn well going to have a coordinate system for all "rooms" such that it would easily calculate who was within a certain radius of any action. I grabbed my old maths book, did the trigonometry, came up with the relevant equations, boiled them down, and then proceeded to implement them in the day's game driver. The game driver had NO support for floating point operations, so I used a few tricks:
- Refer to coordinates in degrees, but increased a couple of orders of magnitude to make fractions of coordinates representable (store a precise value as an acceptably less precise integer). I even scaled it - 800 "degrees" instead of 360 seemed to work best for my needs.
- Implement simple series for some trigonometric calculations - I got sufficiently accurate values with IIRC the first 5 steps in the series
- Where speed really was an issue, I used incomplete trigonometric lookup tables, with some allowance in the functions to approximate results as "x/yths between a[z] and a[z+1]."
- Bingo, one working coordinates scheme to accomodate the physics I wanted to allow for, with absolutely no real numbers in play.
I did not even think of patenting it. Why? Because it only took an afternoon's worth of thought to do all the necessary derivations and calculations, and it was all common sense! Besides, I might have had to have credited my high school math text!The implementation of all the basics in LPC took me probably the better part of a week, and was tested, and worked to an acceptable degree, but that was just coding. It's just a shame I went on to start University before I could release it and give it momentum to grow.
[rant]Grow up Microsoft, and patent something that a high school student can't do.[/rant]
Unfortunately, if we don't have complex "Don't start with a number, the new one must not be similar to the last, do this, don't do that" rules, users will tend to take the easy way out and use "password" if given the option. It seems today that the only way to ensure something random is to reduce the number of allowable permutations. Dictionary cracks become meaningless when the user has no statistical preference for leaning on dictionary words. Given the choice, I would just as likely use "A2jj*Z,L" as "dictionary" for a password, but Joe Average goes and spoils it...